Welcome back everyone, I'm Jordan Geisigee and this is The Limiting Factor. In the last video, I showed that the energy density of the Generation 2 4680 has increased by 11.5% from the Generation 1 4680 and is now highly competitive with the best third-party alternatives that Tesla uses in their vehicles. The question is, how did they achieve such a large increase in energy density? How does that align with my past predictions? And what does that mean for future energy density increases? To answer that question, today I'm going to walk you through the chemistry and design of the Cyber Cell thanks to a teardown and testing by UC San Diego. Before we begin, a special thanks to my Patreon supporters, YouTube members and Twitter subscribers, as well as RebellionAir.com. They specialize in helping investors manage concentrated positions. RebellionAir can help with covered calls, risk management and creating a money master plan from your financial first principles. For this video, there's a number of additional people to thank here. It takes a community effort to make a video like this possible. First, thanks to Monroan Associates for donating the battery cell for the teardown. Without their generosity, the video wouldn't have been possible. Second, thanks to Brian White of Futuraza, who drove the battery cell from the Tesla Owners Club of Michigan Summer meetup all the way to UC San Diego for testing. Third, thanks to Shirley Monk, who organized funding for the teardown as a professor at the University of Chicago and an adjunct professor at UC San Diego. Lastly, thanks to Wei Kung Li and Wei Liang Yao, who tore down the 4680 cell, as well as did the testing to provide the information we're going to look at today.
To kick things off, two years ago, when I released the results from the testing from the Generation 1 4680 cell, I showed that the thickness of the cell can was roughly double the thickness of a 21 70 cell can. By my back of the napkin math, I estimated that reducing the thickness of the 4680 cell can, from about 0.6 millimeters to 0.3 millimeters to match the 21 70 cell, would increase its energy density by up to 24Wh per kilogram. The Generation 2 4680 cell, as I showed in the 4680 teardown video last month, was around 0.4 millimeters thick as compared to the 0.6 millimeters thick of the Generation 1 4680 cell can, which was a reduction in thickness of 0.2 millimeters. That is only about 2 thirds of the 0.3 millimeters I was hoping for. However, I've since been in contact with James Ma of Kaishan Road. James is based out of China and does comprehensive battery teardown reports. His measurements, which were more thorough, showed that the Generation 2 4680 cell can, listed here as Generation 1.5, had a thickness of 0.35 millimeters. That would put the thickness reduction from the Generation 1 to the Generation 2 4680 cell can, at closer to 0.25 millimeters, or about 83% of the 0.3 millimeters that I suggested two years ago.
83% of 24Wh per kilogram would mean a 20Wh per kilogram increase in energy density. If the Generation 1 4680 was 244Wh per kilogram, that would mean a hypothetical energy density increase to 264Wh per kilogram simply by making the cell can thinner. However, as we saw in the previous video, the measured energy density of the Generation 2 4680 is 272Wh per kilogram, which means there's clearly also an improvement to the chemistry of the battery cell. With that in mind, let's take a look at the cathode and anode chemistry of the Generation 2 4680. In the past two years, I've collaborated with UC San Diego to tear down three 4680 battery cells. The results of each are shown on screen. The first and second cells were both Generation 1 4680 battery cells. And the third cell is of course the Generation 2 4680. The Generation 1 4680 cells were roughly the same in terms of composition. 81% nickel, 12% cobalt, and 7% manganese. This would be loosely classified as an NMC 811 cathode chemistry, corresponding to 80% nickel, 10% manganese, and 10% cobalt. The Generation 2 4680 cell is 91% nickel, 5% cobalt, and 4% manganese, which would be loosely classified as an NMC 955 chemistry, correlating to 90% nickel, 5% manganese, and 5% cobalt.
There's two significant points about that. First, Tesla's now using an advanced cathode chemistry in the 4680. The higher the nickel content of a lithium ion cathode material, the greater the energy density. That's because, as we can see on screen, for a given voltage, increased nickel content results in increased capacity. But increasing the nickel content also reduces the stability of the cathode, which has solutions but makes high nickel cathodes more difficult to master. NMC 811 type cathode materials, which are 80% nickel, hit the market about 5-6 years ago, but now they're common. NMC 955 type cathode materials, which are 90% nickel, started to hit the market about 2-3 years ago. As far as I'm aware, there's nothing on the market that has a greater percentage of nickel and a lower percentage of cobalt.
The second significant point about the NMC 955 cathode is that it validates a post on X by Joe Techbier. In January of this year, Joe reported that Tesla just finished the process of transitioning to NMC 955, which matches what the analysis from UC San Diego shows. If you don't follow Joe on YouTube or X, I'd recommend it. He's proven to be a reliable source of original information and news in the Tesla community. In the same post, Joe stated that Tesla was also testing an NMC 973 cathode, which they may use in 4680 cells by the end of the year. NMC 973 would likely have a minimal impact on the energy density because it would still be 90% nickel, but it would decrease the cobalt content to about 3%. That type of cathode would put Tesla at the cutting edge of cathode material science and reduce concerns around cobalt mining.
Looking back on track, how much does the NMC 955 cathode increase the energy density of the 4680 cell? In order to get a ballpark estimate, we need to first take a look at the anode chemistry. The slide on screen shows that just like the generation 1-4680, the generation 2-4680 doesn't contain silicon, which is often used to boost the energy density of lithium ion battery cells. Instead, the anode is effectively pure graphite, or to be more accurate, if there is silicon in the 4680, it's below the detectable limit, and so it would be minuscule and have no real impact on energy density. Since the anodes in the generation 1 and generation 2-4680 both use pure graphite, on a weight for weight and volume for volume basis, the energy density of the anode in each battery cell should be nearly identical. That piece of information can be used to guess the energy density of the cathode material. How? Let's start by comparing the thickness of the anode material for each 4680 that UC San Diego has tested.
Both of the generation 1-4680 cells had a total anode thickness of 250 microns, whereas for the generation 2-4680 it was 240 microns. It's relevant because in a lithium ion battery, the lithium storage capacity of the anode has to roughly match the lithium storage capacity of the cathode. The cathode is where the lithium ions start out. If the cathode contains less lithium than the anode, then there would end up being empty spaces in the anode when it's charged, meaning dead weight and wasted space. If the cathode contains too much lithium, when the battery is charged it would overwhelm the anode and deposit a layer of pure lithium, called lithium plating, which would destroy the battery cell. That means that the generation 2-4680 is using the same anode material as the generation 1-4680, but the thickness of the anode has reduced from 250 microns to 240 microns, or about 4%. The cathode should also see a reduction in thickness of about 4%.
But that's not what we see when we look at the cathode. Instead, the thickness of the cathode has gone from 180 microns to 150 microns, which is a reduction in thickness of 17%. That could have occurred in two ways. First, the cathode material may have experienced greater compression or calendaring during the manufacturing process, which reduced its porosity leading to a thinner coating. This is unlikely because I doubt Tesla could further optimize the porosity by another several percent, let alone another 17%. Second, the cathode material is now higher energy density, which we know should be the case because the cathode is now using more nickel. With that in mind, let's assume the thinner cathode in the generation 2-4680 is all due to a higher energy density cathode material, rather than lower porosity. If the cathode of the 4680 should have been 4% thinner based on the anode, but it ended up being 17% thinner, that would mean relative to the generation 1-4680, the cathode in the generation 2-4680 would have an energy density that's roughly 13% greater. Would that mean 13% greater energy density at the cell level? No, because the cathode of a lithium ion battery cell is about 31% of the total weight of the battery cell. So if we multiply 13% by 31%, that's roughly a 4% energy density increase. 4% times the 244 watt-hour per kilogram baseline of the generation 1-4680 is 10 watt-hours per kilogram. If we add that to the 264 watt-hour per kilogram hypothetical cell energy density from the thinner cell can, that would be a total hypothetical energy density of 274 watt-hours per kilogram, which is within 2 watt-hours per kilogram of the energy density that Monroan associates measured. Given that this is all back of the napkin mass and taking into account measurement errors, the hypothetical energy density of 274 watt-hours per kilogram based on the improvements we saw on the teardown and the measured energy density of 272 watt-hours per kilogram are effectively the same. That means the 11.5% energy density increase at the cell level can be fully explained by a thinner cell can and higher energy density cathode material.
Now that we've looked at the energy density, let's look at a few other facets of the generation 2-4680. First, does the generation 2-4680 use a dry process on both the cathode and anode? No, UC San Diego found that once again, just the anode uses a dry manufacturing process, but we already had advice from Tesla to that effect. As I reported last month in the Q2 earnings call, Tesla said that they intend to start mass production of the fully dry 4680 in the fourth quarter. By implication, that means just like the generation 1-4680, any generation 2 cells produced earlier this year would have used the dry process on just the anode. As usual, I'll continue to keep you updated on the 4680 ramp after each earnings call and the significance of any developments. Next up, resistance. UC San Diego wasn't able to test the resistance of the generation 2-4680 because it was damaged from being discharged to .44 volts and therefore wouldn't provide reliable data. However, I expect the resistance in the generation 2-4680 to be lower than the generation 1-4680 for three reasons. First, as I showed in the teardown video, the copper anode foils of the generation 2 cell are welded directly to the bottom lid of the battery cell. That's as opposed to the generation 1 cell where the copper foils are welded to a current collector which in turn was attached to the cell can. That is, there were more interfaces which may have increased to the electronic resistance.
Second, additionally, the aluminum current collector for the cathode in the generation 2 cell is a solid disk rather than slotted like the generation 1-4680 which means more metal which also potentially means lower electronic resistance. With that said, I expect the improvements to resistance from the anode and cathode terminals to have a trivial effect on the total resistance of the battery cell. Why? Because most of the resistance in a battery cell is due to ionic resistance rather than electronic resistance.
That brings us to the third reason why I expect the resistance of the new 4680 to be lower. As I showed earlier, the cathode and anode are thinner. That in turn lowers ionic resistance which can significantly reduce heat generation when the cell is charging and discharging. It won't be a massive improvement but certainly significant which may have an impact on charging speed.
On that note, a number of people online have been saying that the charging speed of the cyber truck is terrible. As shown on screen, the charging speed of the cyber truck is roughly middle of the pack when compared to other electric pickup trucks. That is not great but also not terrible. However, I expect the cyber truck and 4680 to achieve faster charging speeds in the future.
That's for three reasons. First, earlier this year Tesla said they'll be pushing a software update to increase the charging speed of the cyber truck. As far as I'm aware, that hasn't happened yet and the update is already about three months overdue. But that's par for the course for Tesla software updates. Second, Tesla is working on uncorking the maximum charger output of the V4 Superchargers which would primarily benefit the cyber truck because it has the largest battery pack in Tesla's passenger vehicle fleet and therefore has the greatest capacity to absorb those extra kilowatts. Third, in the longer term, Tesla has a number of options to increase the charging speed of their battery cells from a design and chemistry perspective.
On that note, let's move on to the final topic of the video, which is the potential technical improvements and performance we might expect from the 4680 in the future. In past videos, I showed this image where I suggested that the generation 24680 would hit 268Wh per kilogram and the generation 34680 would have a slightly better design and include a higher nickel cathode. Installing in an energy density of 280Wh per kilogram, it looks like Tesla rolled the improved cathode I predicted for generation 3 into generation 2 and achieved an energy density of 272Wh per kilogram.
There may still be some room to improve the cell design and the cathode, which could take it closer to my estimate of 280Wh per kilogram, but let's assume 272Wh per kilogram is the current maximum energy density without resorting to added silicon in the anode, asymmetric lamination of the electrodes, or lithium doping. Even at 272Wh per kilogram, adding an industry standard amount of silicon to the battery cell could still potentially bring the 4680 up to 300Wh per kilogram.
That's because when I made this table, I was purposely conservative. That means my estimates of 315Wh per kilogram for asymmetric lamination and 330Wh per kilogram for lithium doping also still stand. The question is, what kind of timeframes could we be looking at for those energy density increases? It's impossible to say without inside knowledge, but here's my thoughts. For increased silicon, I'd say about 1 to 2 years.
I'd only caution that I wouldn't be surprised if Tesla started with small amounts of silicon and slowly increased over time. So we may not see the leap to 300Wh per kilogram in one generation. As a side benefit, increasing the silicon content of the anode would likely increase the charge rate of the 4680. If you'd like to know why, check out my video on how silicon affects charging speed.
Although the predictions in the video haven't panned out yet, the technical reasoning is solid. As for asymmetric lamination, the time frame is a total wildcard. However, in the post where Joe Techmyer accurately reported the NMC955 cathode, he also said that Tesla was already in the process of trialing an asymmetric electrode. That could mean a rollout as soon as the generation 3.4680. Or it could mean that Tesla is in a very early trial phase, and still working out the fundamentals of asymmetric lamination, which could mean it's still years away.
Asymmetric lamination can be used to both increase energy density and increase charging speed. So if Tesla does implement that technology, it's really a product decision as to how they split the benefits. If you'd like to know more about that, I've linked a video above. Lithium doping is a technology that I don't expect anytime soon because it involves working with pure lithium metal, which is volatile. But I'm hoping to see it this side of 2030. Once again, if you'd like to know more about that, watch my Tesla Lithium Doping Patent video, which is linked above.
In summary, thanks to a thinner cell can and higher energy density cathode material, the Tesla 4680 is now on par with the best high nickel battery cells on the market. Tesla has a lot of options to continue to improve the battery cell. With that said, it's best to keep our eye on the prize, which is mass production of the fully dry 4680, which is expected to start next quarter. If Tesla continues at their current pace of improvement, in terms of both cost and performance, they could gain a significant lead on the entire battery industry in as little as the next two years. It all depends on what other manufacturers unveil in the next couple of years. As usual, I'll keep you updated on major developments in my regular videos, and after each Tesla quarterly earnings call.
Before I close the video, I'd like to say thanks again to all the people who made this video possible, including Rebellionair, Sandy Monroe, Brian White, Shirley Mung, Wei Kang Li, and Wei Liang Yao. If you enjoyed this video, please consider supporting the channel by using the links in the description. Also consider following me on X. I often use X as a testbed for sharing ideas, and X subscribers like my Patreon supporters generally get access to my videos a week early.
在结束视频之前,我想再次感谢所有帮助制作这段视频的人,包括Rebellionair、Sandy Monroe、Brian White、Shirley Mung、Wei Kang Li和Wei Liang Yao。如果你喜欢这个视频,请考虑通过视频描述中的链接支持这个频道。同时也可以考虑关注我的X账号。我经常在X平台上测试分享一些想法,X的订阅者和我的Patreon支持者一样,一般能够提前一周观看到我的视频。
On that note, a special thanks to my YouTube members, X subscribers, and all the other patrons listed in the credits. I appreciate all of your support, and thanks for tuning in.